基于莱特准则的椭圆拟合优化算法improvedellipse-计算机应用.pdfVIP

Journal of Computer Applications ISSN 1001-9081 2016-12-10 计算机应用,2016,36(12) CODEN JYIIDU 基于莱特准则的椭圆拟合优化算法 * 曹俊丽 ，李居峰 (上海大学 机电工程与自动化学院，上海市 200072) (*曹俊丽caojunli1009@163.com) 摘 要：普遍使用的代数距离最小的最小二乘椭圆拟合算法简单，易实现，但对样本点无选择，导致拟合结果易受误差点 影响，拟合不准确。针对此特性，提出了一种基于莱特准则的椭圆拟合优化算法。首先，由代数距离最小的最小二乘法对待 拟合曲线进行椭圆拟合；其次，将待拟合曲线上的点距离最小二乘法拟合的椭圆的代数距离作为样本点集，在验证该样本点 集服从正态分布的情况下，采用莱特准则，将样本点中值大于|3σ|的点判定为野值并剔除，进行多次拟合，直至样本点中无野 值。最后，得到椭圆最优拟合结果。仿真实验结果表明，优化算法的拟合误差在 1.0%以下，相比同条件下的最小二乘法，其 拟合精度至少提高2%。优化算法的仿真结果与其在香烟圆度在线检测中的实际应用验证了此算法的有效性。 关键词: 莱特准则；椭圆拟合；最小二乘法；圆度检测；视觉检测系统 中图分类号:O241.5 文献标志码: A Improved ellipse fitting algorithm based on Letts Criterion CAO Junli*, LI Jufeng ( School of Mechatronics Engineering and Automation, Shanghai University, Shanghai 200072, China) Abstract: The Least square ellipse fitting algorithm based on minimum algebraic distance was simple and easy to implement, but it had no choice to the sample points, which lead to the error of the fitting result, and then the fitting is not accurate. According to this case, a new ellipse fitting algorithm based on Letts Criterion was proposed to overcome the shortage of Least Square alg orithm. First, the ellipse was fitted using the Least square ellipse fitting algorithm based on minimum algebraic distance . Then, the algebraic distance between the points of ellipse fitted by Least Square algorithm and the point on the fitting curve was set as fitting point set. After verifying the point set was in normal distribution, the points which were greater than |3σ| were eliminated using Letts Criterion. Then the steps were repeated above